Intelligent knowledge recommending approach for new product development based on workflow context matching

The variety of product types/specifications in discrete manufacturing enterprises makes new product development tasks real tough work. Therefore, it is a common strategy for workers to refer to similar outcomes (e.g. the product drawings and work instructions) of former new product development tasks. In order to discover similar historic outcome, this article presents an intelligent approach to measure the cohesion between workflow contexts in process-aware information systems and exploit it for runtime task knowledge recommendation. The measure of context similarity is preceded by (1) modeling the task context with ontology theory and (2) using the ontology matching algorithms to evaluate the similarities between context ontology entities of different tasks. Specifically, the term frequency-inverse document frequency approach is utilized to compute the context cohesion between current task and historic ones, and the tasks with the highest similarity will be recommended to the task executors, along with their outcomes. Comparative evaluation is performed using term frequency-inverse document frequency, Levenshtein, and Affine Gaps, and results demonstrate that the proposed approach achieves good precision and recall and is efficient in task knowledge recommendation.